Cigarette smoking is a major contributor to cancer, heartdisease, stroke, and lung disease and is the leading cause of preventable mortality in the United States and worldwide. Genome-wide association studies (GWAS) of nicotine dependence (ND) and other smoking phenotypes have unequivocally identified associations with single nucleotide polymorphisms (SNPs) spanning the nicotinic acetylcholine receptor (nAChR) genes on chromosome 15q25 (CHRNA5-CHRNA3-CHRNB4) and on chromosome 8p11 (CHRNB3- CHRNA6). The functional variants alter cessation treatment response and are independently associated with health consequences of smoking. However, they explain <10% of the phenotypic variability for smoking. In this study, we will move beyond standard GWAS, conducting the first study to apply the joint 2 degree-of- freedom (2df) statistical method using gene-gene interaction. It will be the largest genome-wide study of ND. The newly developed joint 2df method simultaneously considers SNP main and interactive effects but does not focus on detecting interaction per se. Instead, the method leverages interaction to increase power of the SNP association test. It has been used to successfully identify SNP associations missed in standard GWAS for other complex traits when accounting for gene-environment interactions. Here, we will apply the joint 2df method to simultaneously consider SNP main associations and their interactions with the established nAChR variants. We capitalize on existing GWAS data from our Collaborative Genetic Study of ND (COGEND), deCODE Genetics, and other studies available via the Database of Genotypes and Phenotypes (dbGaP) with Fagerstrom Test for Nicotine Dependence (FTND) data: collective size of >23,000 participants. In Aim1, we will obtain genome-wide SNP genotype data from 14 study samples and harmonize their FTND phenotype data. Using COGEND and deCODE Genetics (total N=10,319 Caucasians), we will conduct genome-wide joint 2df meta-analyses of ND, in parallel, accounting for an interaction with either CHRNA5- CHRNA3-CHRNB4 (in Aim 2) or CHRNB3-CHRNA6 (in Aim 3). The top-ranking SNPs from the joint 2df meta-analyses will be prioritized using bioinformatics analyses to evaluate their annotation, functionality, and regulatory potential. In Aim 4a, we will select up to 100 top-ranking SNPs for replication testing in seven independent Caucasian samples from dbGaP (total N=8,913). Replicated SNPs from Aim 4a will be further tested for association with ND in Aim 4b using two of our African American samples and three dbGaP samples (total N=4,145). Our prior GWAS of smoking and others have identified robust, functionally important variants that also contribute to cessation treatment response and smoking-related diseases (e.g., lung cancer). This study leverages these early findings and a new statistical method to discover novel variants contributing to ND, which may similarly add to our understanding of smoking cessation and smoking's health consequences.